[omniORB] Installation and Configuration on RedHat 6.0

Latham, Steve SLatham@acdnj.itt.com
Tue, 14 Sep 1999 10:47:08 -0500


I've compiled and run omniORB2 2.8 and all examples on Redhat 6.0 with no
problems

-----Original Message-----
From: Sai-Lai Lo [mailto:S.Lo@uk.research.att.com]
Sent: Tuesday, September 14, 1999 5:20 AM
To: mike@twinpeaks.prc.com
Cc: omniorb-list@uk.research.att.com
Subject: Re: [omniORB] Installation and Configuration on RedHat 6.0


>>>>> Mike Ladwig writes:

> Sai-Lai Lo wrote:
>> One thing you should watch out for with RH6 is that it comes with
>> glibc-2.1. It runs programs compiled and linked  with glibc-2.0, i.e. RH
>> 5.2 fine. BUT if you compile on RH6 and link with a library, e.g.
omniORB2
>> runtime, that was linked with glibc-2.0, the executable won't work.

> I thought I had read on this mailing list (and from you in private email)
> that RH 6.0 (with glibc 2.1) wasn't working well with omniORB, even when
the
> entire system is recompiled.  Did I misunderstand, or is this something
that
> has been improved with the omniORB 2.8 releases?

Well, I've reported on this list that we are staying with RH 5.2 because of
the incompatibility we've found. Briefly, if you mix libraries and object
files compiled under one environment with another, you end up with an
executable that just core dump. 

I've also reported problems with iostream library of glibc-2.0 which causes
core dump when multiple threads try to write to cerr at the same time. 
Helmut Swaczinna noted that the gcc iostream library is indeed non-threaded
safe. That is not a problem because the ORB internal do not use cerr. It is
only in my testing code that cerr is used. I would also assume the same
applies to RH 6 (glibc-2.1).

It is also known that the Suse 6.1 distribution, which comes with glibc-2.1,
has a egcs-1.1.2 compiler that were not compiled with
--enable-threads. So the code generated by it core dump when you have
multiple threads raising an exception.

I think the only known problem with RH 6, assume that you have everything
built from scratch, is the loader problem you have reported earlier. It
would be useful if you can dig deeper to see what is wrong. I doubt it is a
bug in the code. More likely a runtime or compiler problem. Finally, I've
not seen any positive confirmation that the egcs-1.1.2 that comes with RH 6
is compiled with --enable-threads. If there is any report of random crash
when running the example programs, the compiler will the prime suspect.

Sai-Lai



-- 
Sai-Lai Lo                                   S.Lo@uk.research.att.com
AT&T Laboratories Cambridge           WWW:   http://www.uk.research.att.com 
24a Trumpington Street                Tel:   +44 1223 343000
Cambridge CB2 1QA                     Fax:   +44 1223 313542
ENGLAND